Reinforcement learning in the robocup-soccer keepaway

By Javier García, Fernando Fernández, Manuela M. Veloso. 2007

396 citations recorded by the source corpus.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

A Deep Hierarchical Approach to Lifelong Learning in Minecraft (2016, background): “Many of the tasks that are encountered by an agent in a lifelong learning setting can be naturally decomposed into skill hierarchies [Stone et al. , 2000; Stone et al. , 2005; Bai et al. , 2015].”
Integrating reinforcement learning with human demonstrations of varying ability (2011, methodology): “As a way of incorporating domain knowledge, the learners choose not from the simula-tor’s primitive actions but from a set of higher-level macro-actions implemented as part of the player [24].”
Transfer via inter-task mappings in policy search reinforcement learning (2007, background): “Likewise, the action Pass to closest keeper is analogous in both tasks, as is Pass to second closest keeper.”
Planning with macro-actions in decentralized POMDPs (2014, background): “, robots with actions that execute predefined controllers) [29], more efficient planning [30], skill transfer [16], and skill-specific abstractions [12].”
Sketch-Based Linear Value Function Approximation (2012, methodology): “In continuous-state RL domains, tile coding [23] is a canonical example of exhaustive feature generation; tile coding has been successfully applied to benchmark domains [22], to learn to play keepaway soccer [20], in multiagent robot learning [4], to train bipedal robots to walk [18, 24] and to lear”
TAMER: Training an Agent Manually via Evaluative Reinforcement (2008, background): “Reinforcement learning has recently made great strides in terms of applicability to complex sequential decision making tasks [1], [2], [3].”
Experience Replay for Real-Time Reinforcement Learning Control (2012, background): “1) Robotic Goalkeeping Problem: Robotic soccer is a popular benchmark problem for RL [22], [30].”
Half Field Offense: An Environment for Multiagent Learning and Ad Hoc Teamwork (2016, background): “Keepaway created interest , sparked research, and has served for many years as a testbed for new AI algorithms [19] [25] [26].”
Batch reinforcement learning in a complex domain (2007, methodology): “[11] (depicted in Figure 1) involve distances and angles between the other players and K1.”
Keepaway Soccer: From Machine Learning Testbed to Benchmark (2005, methodology): “The keepers learn their task using episodic SMDP Sarsa(λ) [17] [16], a well-understood temporal difference algorithm and naturally fit into the SMDPagent interface.”